Code Documentation

Modules

DirichletMagmaMix

A python package for modelling partial mantle melt aggregation using the Dirichlet distribution as described by Rudge et al. (2013).

Developed and maintained by Simon Matthews (simonm@hi.is).

The module requires a set of primary melt compositions to be generated elsewhere. At the present time the module can only important results generated by the pyMelt library (though an older version of the code is compatible with alphaMELTS).

exception DirichletMagmaMix.Error[source]

Base class for exceptions in this module.

exception DirichletMagmaMix.InputError(message)[source]

Exception raised for errors in the input.

Attributes:

expression – input expression in which the error occurred message – explanation of the error

class DirichletMagmaMix.MixedMelts(melts)[source]

Results from a dirichlet mixing calculation.

class DirichletMagmaMix.UnmixedMelts(melts, pressure=None, mass=None, weights=None, isotopes=[], mass_warning=True, lithology_names=[], **kwargs)[source]

Contains all the melts (from all lithologies) and information about how they should be weighted in the mixtures.

Parameters

  • melts (pandas.DataFrame) – All of the melts from the melting region. Columns should only be chemical composition.

  • weighting (str, pandas.Series, or None, default: None) – How to weight the melts in the mixtures (in addition to their mass). In built options: - ‘triangle’, a triangular melting region for a spreading-centre - ‘truncated_triangle’, a triangular melting region missing the

    top or bottom. Specify by passing arguments ‘shape_top’ and ‘shape_bottom’ (pressure in GPa).

    Alternatively, supply a pandas Series with the numerical weighting values. If None, an equal weighting will be applied.

  • isotopes (list of dicts, default: []) – For any isotope ratios in the melts DataFrame a dict should be provided with the following keys: - ‘ratio’: the name of the ratio in the DataFrame, e.g., ‘Nd143Nd144’. - ‘element’: the element the isotopes are associated with, e.g., ‘Nd’.

    This element must be in the DataFrame.

  • mass_warning (bool, default: True) – Warn if there is no mass information provided.

  • lithology_names (list of strings, default: []) – The names of the lithologies contributing to the melting column. Used to preserve tracer variables, for example during fractional crystallisation.

correlation()[source]

Calculates the correlation coefficient between elements. This is independent of the mixing parameter.

Returns

DataFrame containing the correlations with elements as the indexes and columns.

Return type

pandas.DataFrame

gen_cmc_melts(samples, max_crystallisation, mixing_min, mixing_max, partition_coefficient=None)[source]

Generate melts produced by concurrent mixing and crystallisation, using the model defined in Rudge et. al. (2003).

Parameters

  • samples (int) – Number of mixed melts to produce.

  • max_crystallisation (float) – Maximum fraction of crystallisation. Denoted by Chi in Rudge et. al. (2003)

  • mixing_min (float or int) – Minimum degree of melt mixing

  • mixing_max (float or int) – Maximum degree of melt mixing

  • partition_coefficient (float, pandas.Series or None, default: None) – The partition coefficient for each element during crystallisation. If None, all elements will be treated as being perfectly incompatible.

Returns

The randomly mixed melts.

Return type

MixedMelts

gen_dirichlet_parameters(N)[source]

Generate dirichlet parameters given the mixing parameter passed to function, and the weighting variables contained in UnmixedMelts object.

Parameters

N (float or int) – Mixing parameter. Must be >1.

Returns

Dirichlet parameters

Return type

pandas.Series

gen_mixed_melts(mixing_parameter, samples=1)[source]

Generate mixed melts with mixing parameters randomly selected from the dirichlet distribution defined by the mixing parameter.

Parameters

  • mixing_parameter (float or int) – Mixing parameter. Must be >=1.

  • samples (int, default: 1) – Number of mixed melts to return.

Returns

The randomly mixed melts.

Return type

MixedMelts

gen_mixing_weights(alpha)[source]

Generate a random distribution of mixing proportions.

Parameters

alpha (pandas.Series) – The dirichlet parameters, probably calculated using the gen_alpha method.

Returns

Mixing weights for one random draw from the dirichlet distribution.

Return type

pandas.Series

homogenise(pressure)[source]

Homogenise melts produced below a certain pressure.

Parameters

pressure (float) – Pressure at which to homogenise melts generated beneath.

mean_composition()[source]

Calculates the mean composition of the melts in the melting column, according to the defined masses and weighting.

Returns

The mean composition

Return type

pandas.Series

variance(mixing_parameter=1)[source]

Calculate the variance for elements at a specified degree of mixing. Uses the equation given in the appendix of Rudge et al.

Parameters

mixing_parameter (float or int) – Mixing parameter.

Returns

Series with element names as index.

Return type

pandas.Series

DirichletMagmaMix.import_pyMelt_geoSetting(geoSetting, weights=None, isotopes={}, **kwargs)[source]

Imports a pyMelt geoSetting.

Parameters

  • geoSetting (pyMelt.geoSetting) – The geoSetting class containing the melts to be mixed.

  • weights (pandas.Series, string, or None, default: None) – Specify weighting for the melts (see UnmixedMelts docstring for more information). If None and the geoSetting is a spreadingCentre, a triangular weighting will be applied, otherwise the melts will have equal weighting.

  • isotopes (dict containing lists of dicts, default: {}) – The isotope ratios to assign to the melts. The keys of the first dictionary should be the lithology names, then the list consists of dictionaries containing: - ratio, e.g., ‘Nd143Nd144’ - element, e.g., ‘Nd’ - value, e.g., 0.51230

Returns

The UnmixedMelts object created during import

Return type

UnmixedMelts